Although the steps by which the brain encodes color signals are relatively well understood, surprisingly little is known about the neural computations that decode these signals to yield color percepts. One hint comes from lesions of inferior temporal cortex (IT) in rhesus monkeys that cause a profound loss of color perception. Using fMRI-guided single-unit recording in monkeys, we found a system of IT modules that likely play an important role in representing color and are the probable homologue of color regions in humans. The relationship of the color modules to the rest of IT is poorly understood, as is the link between the neural activity within them and perception. By contrast, perceptual chromatic mechanisms in humans have been studied extensively, giving rise to many hypotheses that could be tested in monkeys with recordings targeted to the color modules. But it is unknown whether the two species possess similar psychophysical chromatic mechanisms, imposing a critical barrier to progress in the field. Our overarching aim is to overcome this obstacle and to uncover the neural computations for color perception using fMRI-guided microelectrode recording of IT in trained monkeys engaged in color tasks. The central hypothesis, based on substantial preliminary data, is that monkeys possess both cardinal and higher-order chromatic mechanisms similar to humans; and that these mechanisms are implemented by essential transformations that culminate in the representation of perceptual color space within specialized IT color modules. The proposal consists of three aims. First, we will modify for monkeys the classic paradigm used to reveal psychophysical chromatic mechanisms in humans. Second, we will use fMRI to determine the extent to which the modules are specialized for color and functionally connected. Third, we will address for the first time the link from color-cell activityin posterior IT to color-discrimination behavior, using signal-detection theory to compute choice-probability estimates from neural recordings obtained while animals perform a color task. Achieving the proposal's aims will facilitate the next logical steps toward our long-term goal of understanding the computations underlying color perception, including the construction of a physiologically plausible model of color. The research will inform general principles of neural circuit function, which is critical for understanding the etiology, diagnosis and treatment of mental illness and neurological disease; and will guide treatments for visual impairment that bypass the eye and exploit those cortical computations that survive the primary insult in blindness.